In total, 221 patients who underwent surgery for unstable pelvic ring injuries from May 2012 to May 2021 were reviewed. The inclusion criteria for these patients were as follows: (1) fixation of two ipsilateral IS S1 screws (7.3 or 7.0 mm); (2) fixation procedure performed percutaneously under two-dimensional fluoroscopic guidance; (3) at least a 1-year postoperative follow-up; and (4) available medical charts and radiographs to review exact data. Of the 221 patients, 38 patients were included in the analysis. Table 1 shows the included patients’ demographic data, and Table 2 shows the fixation constructs of the included patients.

Before surgery, we measured the inlet angle, which is tangent to the anterior cortex of S1, and the outlet angle, which is perpendicular to the midline of the trapezoidal S1 body, using a midline sagittal computed tomography (CT) image (Fig. 1A). Based on the measured angles and fluoroscopic findings, we accurately determined the angle of the inlet and outlet views. The inlet view visualized the anterior cortex of the S1 body and spinal canal, whereas the outlet view visualized the sacral foramen (Fig. 2A). Lines perpendicular to the determined inlet and outlet angles were drawn on the patient along with the line of iliac cortical density (ICD), which could be confirmed in the true lateral view (Fig. 1B). Surgery was then performed on a radiolucent operating table, and the patient was prepared in the supine position. The entry of the first guide pin was made anterior in S1 and inferior to ICD. The height of the entry and coronal direction of the guide pin were determined via the outlet view, and the anterior-posterior position and direction were determined through the inlet view (Fig. 2B, C). The drawn perpendicular line of the determined inlet and outlet angles helped modify the position of the guide pin entry and direction of the guide pin. For example, if the guide pin position or direction needed modification in the outlet view, moving the pin along the line drawn perpendicular to the outlet angle could modify the position or direction of the pin effectively in the outlet view. After making the accurate position of entry and direction, we advanced the guide pin. The guide pin was advanced cranial to the S1 foramen in the outlet view and near the anterior cortex, but within it, in the inlet view (Fig. 2D, E). After passing the S1 foramen, the true lateral view was used to confirm that the ICD had not been penetrated, and the guide pin was then advanced to the S1 body (Fig. 2F). Finally, reaming was performed, and a cannulated screw with a washer was inserted via the guide pin. When tightening the screws, the obturator inlet view was used to confirm the narrowing of the sacroiliac joint gap, thereby confirming the precise state of the screw head and washer to prevent excessive intrusion (Fig. 2G–I). The second unilateral IS screw was fixed whenever possible; however, if the safe corridor was not sufficiently wide and pelvic dysmorphism was observed, the second screw was not used. The guide pin of the second screw was inserted; the guide pin of the first screw was not removed because it helped to guide the direction of the second guide pin (Fig. 3A). The second guide pin was positioned cranial to the first screw in the outlet view and positioned in a manner that did not penetrate the anterior cortex or the spinal canal in the inlet view and ICD in the lateral view (Fig. 3A–C). As a result, the second guide pin usually overlapped with or was slightly posterior to the first screw. The interval between the first screw and the second guide pin was at least half of the screw diameter to prevent collision between the screws. After confirming the adequate position of the second guide pin, the second screw was inserted (Fig. 3D–F). The type of screws was selected based on the need for compression. If compression was not required or overcompression needed to be avoided, a fully threaded screw was fixed. If compression was required, a partially threaded screw was fixed. Moreover, if additional compression was required after the first partially threaded screw fixation, the second screw was also selected for the partially threaded screw. A washer was routinely applied.

All patients were instructed to perform a hip joint range of motion exercise and quadriceps strengthening exercises immediately after surgery. Furthermore, “toe-touch” weight-bearing to partial weight-bearing immediately after surgery was initially advised to patients with vertically stable injuries. However, patients with vertically unstable injuries started partial weight-bearing 4 to 6 weeks postoperatively. Thereafter, we progressively increased the amount of weight-bearing according to patients’ tolerance. The patients were followed up at 2 weeks, 6 weeks, 3 months, 6 months, and 1 year postoperatively.

The procedure time was evaluated using fluoroscopic data. Because all procedures were performed under fluoroscopic guidance, we could determine the exact start and end times of the procedure using fluoroscopic data. If surgery involved only percutaneous two ipsilateral S1 screw fixation with intraoperative transfusion, blood loss was evaluated using hemoglobin (Hb) loss between the preoperative and postoperative periods. Intraoperative complications (e.g., guide wire breakage) were also evaluated. Bone union was defined as the absence of a fracture line. In addition, the time to union was evaluated. Moreover, we measured radiological outcomes using the method of Matta and Tornetta [8]; reductions were graded as excellent (<4 mm), good (5–10 mm), fair (11–20 mm), and poor (>20 mm) using maximal displacement measured on the three standard radiographic views of the pelvis. The bone union and radiographic outcomes were determined by a consensus of two orthopedic trauma fellowship-trained independent observers who were blinded to clinical information. Finally, several postoperative complications, including reduction loss, implant loosening or failure, neurological complication, and infection, were evaluated.